【实验17】不同优化算法的比较分析

发布于:2024-12-20 ⋅ 阅读:(13) ⋅ 点赞:(0)

目录

1 不同优化算法比较分析-2D可视化实验

1.1 优化算法的实验设定(以函数y=x^2为例)

1.2 学习率调整优化策略

1.1.2 AdaGrad算法

1.1.2 RMSprop算法

1.3 梯度估计修正优化策略

1.3.1 动量法

1.3.2 Adam算法

1.4 完整代码 

1.5 函数\frac{1}{20}x^2+y^2 的优化算法比较

2 不同优化算法比较分析-3D可视化实验

2.1 函数3D可视化

2.2 加入不同优化算法,画出轨迹

2.2.1 x[0]^{2}+x[1]^2+x[1]^3+x[0]*x[1]

2.2.2 \frac{1}{20}x^2+y^2

3 复现CS231经典动画

4 参考链接

1 不同优化算法比较分析-2D可视化实验

除了批大小对模型收敛速度的影响外,学习率和梯度估计也是影响神经网络优化的重要因素。

神经网络优化中常用的优化方法也主要是如下两方面的改进,包括:

        学习率调整:通过自适应地调整学习率使得优化更稳定。AdaGrad、RMSprop、AdaDelta算法等。
        梯度估计修正:通过修正每次迭代时估计的梯度方向来加快收敛速度。动量法、Nesterov加速梯度方法等。
        综合学习率调整和梯度估计修正:Adam算法

1.1 优化算法的实验设定(以函数y=x^2为例)

        选择一个二维空间中的凸函数,然后用不同的优化算法来寻找最优解,并可视化梯度下降过程的轨迹。

        将被优化函数实现为OptimizedFunction算子,其forward方法是函数y=x^2的前向计算,backward方法则计算被优化函数对x的偏导。代码实现如下:

# 优化函数
class OptimizedFunction(Op):
    def __init__(self, w):
        super(OptimizedFunction, self).__init__()
        self.w = w
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return torch.matmul(self.w.T, torch.tensor(torch.square(self.params['x']), dtype=torch.float32))

    def backward(self):
        self.grads['x'] = 2 * torch.multiply(self.w.T, self.params['x'])

 训练函数 定义一个简易的训练函数,记录梯度下降过程中每轮的参数x和损失。代码实现如下:

# 训练函数,记录梯度下降过程中每轮的参数x和损失
import copy

def train_f(model, optimizer, x_init, epoch):
    """
    训练函数
    输入:
        - model:被优化函数
        - optimizer:优化器
        - x_init:x初始值
        - epoch:训练回合数
    """
    x = x_init
    all_x = []
    losses = []
    for i in range(epoch):
        all_x.append(copy.copy(x.numpy()))
        loss = model(x)
        losses.append(loss)
        model.backward()
        optimizer.step()
        x = model.params['x']
        print(all_x)
    return torch.tensor(all_x), losses

 可视化函数 定义一个Visualization类,用于绘制xx的更新轨迹。代码实现如下:

# 可视化函数,用于绘制更新轨迹
class Visualization(object):
    def __init__(self):
        """
        初始化可视化类
        """
        # 只画出参数x1和x2在区间[-5, 5]的曲线部分
        x1 = np.arange(-5, 5, 0.1)
        x2 = np.arange(-5, 5, 0.1)
        x1, x2 = np.meshgrid(x1, x2)
        self.init_x = torch.tensor([x1, x2])

    def plot_2d(self, model, x, fig_name):
        """
        可视化参数更新轨迹
        """
        fig, ax = plt.subplots(figsize=(10, 6))
        cp = ax.contourf(self.init_x[0], self.init_x[1], model(self.init_x.transpose(0, 1)),
                         colors=['#e4007f', '#f19ec2', '#e86096', '#eb7aaa', '#f6c8dc', '#f5f5f5', '#000000'])
        c = ax.contour(self.init_x[0], self.init_x[1], model(self.init_x.transpose(0, 1)), colors='black')
        cbar = fig.colorbar(cp)
        ax.plot(x[:, 0], x[:, 1], '-o', color='#000000')
        ax.plot(0, 'r*', markersize=18, color='#fefefe')

        ax.set_xlabel('$x1$')
        ax.set_ylabel('$x2$')

        ax.set_xlim((-2, 5))
        ax.set_ylim((-2, 5))
        plt.savefig(fig_name)

        定义train_and_plot_f函数,调用train_f和Visualization,训练模型并可视化参数更新轨迹。代码实现如下: 

# 训练模型并可视化参数更新轨迹
import numpy as np
def train_and_plot_f(model, optimizer, epoch, fig_name):
    """
    训练模型并可视化参数更新轨迹
    """
    # 设置x的初始值
    x_init = torch.tensor([3, 4], dtype=torch.float32)
    print('x1 initiate: {}, x2 initiate: {}'.format(x_init[0].numpy(), x_init[1].numpy()))
    x, losses = train_f(model, optimizer, x_init, epoch)
    print(x)
    losses = np.array(losses)

    # 展示x1、x2的更新轨迹
    vis = Visualization()
    vis.plot_2d(model, x, fig_name)

        模型训练与可视化 指定函数y=x^2中w的值,实例化被优化函数,通过小批量梯度下降法SGD更新参数,并可视化x的更新轨迹。

#1.SGD
from op import SimpleBatchGD
# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = SimpleBatchGD(init_lr=0.2, model=model)
train_and_plot_f(model, opt, epoch=40, fig_name='opti-vis-para.pdf')
plt.title("SGD")
plt.show()

        图中不同颜色代表f(x1,x2)的值,具体数值可以参考图右侧的对应表,比如深粉色区域代表f(x1,x2)在0~8之间,不同颜色间黑色的曲线是等值线,代表落在该线上的点对应的f(x1,x2)的值都相同。 使用SGD这种算法,每次更新的方向就是在该点的负梯度方向,学习率是人为给定的一个固定值,如果过大就不会收敛,如果过小则收敛速度太慢;而且如果学习率是固定值的话,可能在靠近模型最优点时会发生振荡。

        正确的学习率,应该是在开始的时候较大,这样可以加快模型收敛;越往后学习率应该越小,这样方便收敛到极值点。于是出现了AdaGrad算法。

1.2 学习率调整优化策略

        学习率是神经网络优化时的重要超参数。在梯度下降法中,学习率αα的取值非常关键,如果取值过大就不会收敛,如果过小则收敛速度太慢。

常用的学习率调整方法包括如下几种方法:

        学习率衰减:如分段常数衰减(Piecewise Constant Decay)、余弦衰减(Cosine Decay)等;
        学习率预热:如逐渐预热(Gradual Warmup) 等;
        周期性学习率调整:如循环学习率等;
        自适应调整学习率:如AdaGrad、RMSprop、AdaDelta等。自适应学习率方法可以针对每个参数设置不同的学习率。
        下面我们来详细介绍AdaGrad和RMSprop算法。

1.1.2 AdaGrad算法

#2.AdaGrad
from op import Optimizer

class Adagrad(Optimizer):
    def __init__(self, init_lr, model, epsilon):
        """
        Adagrad 优化器初始化
        输入:
            - init_lr: 初始学习率
            - model:模型,model.params存储模型参数值
            - epsilon:保持数值稳定性而设置的非常小的常数
        """
        super(Adagrad, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.epsilon = epsilon

    def adagrad(self, x, gradient_x, G, init_lr):
        """
        adagrad算法更新参数,G为参数梯度平方的累计值。
        """
        G += gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """
        参数更新
        """
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.adagrad(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)

 2D可视化实验 使用被优化函数展示Adagrad算法的参数更新轨迹。代码实现如下:

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = Adagrad(init_lr=0.5, model=model, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para2.pdf')
plt.title("AdaGrad")
plt.show()

       【分析】从输出结果看,AdaGrad算法在前几个回合更新时参数更新幅度较大,随着回合数增加,学习率逐渐缩小,参数更新幅度逐渐缩小。在AdaGrad算法中,如果某个参数的偏导数累积比较大,其学习率相对较小。相反,如果其偏导数累积较小,其学习率相对较大。但整体随着迭代次数的增加,学习率逐渐缩小。该算法的缺点是在经过一定次数的迭代依然没有找到最优点时,由于这时的学习率已经非常小,很难再继续找到最优点。于是出现了RMSprop算法。

1.1.2 RMSprop算法

#3.RMSprop
class RMSprop(Optimizer):
    def __init__(self, init_lr, model, beta, epsilon):
        """
        RMSprop优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta:衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(RMSprop, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.beta = beta
        self.epsilon = epsilon

    def rmsprop(self, x, gradient_x, G, init_lr):
        """
        rmsprop算法更新参数,G为迭代梯度平方的加权移动平均
        """
        G = self.beta * G + (1 - self.beta) * gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.rmsprop(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)

2D可视化实验 使用被优化函数展示RMSprop算法的参数更新轨迹。代码实现如下: 

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = RMSprop(init_lr=0.1, model=model, beta=0.9, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para3-RMSprop.pdf')
plt.title("RMSprop")
plt.show()

        RMSprop算法用到了指数移动平均,是一种自适应学习率的方法,可以在有些情况下避免 AdaGrad 算法中学习率不断单调下降以至于过早衰减的缺点,但是初始学习率是人为指定的,是个固定值,这样的效果并不好,我们希望他是能够动态变化的。于是出现了AdaDelta算法。

1.3 梯度估计修正优化策略

        除了调整学习率之外,还可以进行梯度估计修正。在小批量梯度下降法中,由于每次迭代的样本具有一定的随机性,因此每次迭代的梯度估计和整个训练集上的最优梯度并不一致。如果每次选取样本数量比较小,损失会呈振荡的方式下降。

        一种有效地缓解梯度估计随机性的方式是通过使用最近一段时间内的平均梯度来代替当前时刻的随机梯度来作为参数更新的方向,从而提高优化速度。

1.3.1 动量法

#4.Momentum
class Momentum(Optimizer):
    def __init__(self, init_lr, model, rho):
        """
        Momentum优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - rho:动量因子
        """
        super(Momentum, self).__init__(init_lr=init_lr, model=model)
        self.delta_x = {}
        for key in self.model.params.keys():
            self.delta_x[key] = 0
        self.rho = rho

    def momentum(self, x, gradient_x, delta_x, init_lr):
        """
        momentum算法更新参数,delta_x为梯度的加权移动平均
        """
        delta_x = self.rho * delta_x - init_lr * gradient_x
        x += delta_x
        return x, delta_x

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.delta_x[key] = self.momentum(self.model.params[key],
                                                                      self.model.grads[key],
                                                                      self.delta_x[key],
                                                                      self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = Momentum(init_lr=0.1, model=model, rho=0.9)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para4-Momentum.pdf')
plt.title("Momentum")
plt.show()

【分析】:从输出结果看,在模型训练初期,梯度方向比较一致,参数更新幅度逐渐增大,起加速作用;在迭代后期,参数更新幅度减小,在收敛值附近振荡。 

1.3.2 Adam算法

#5.Adam
class Adam(Optimizer):
    def __init__(self, init_lr, model, beta1, beta2, epsilon):
        """
        Adam优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta1, beta2:移动平均的衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(Adam, self).__init__(init_lr=init_lr, model=model)
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.M, self.G = {}, {}
        for key in self.model.params.keys():
            self.M[key] = 0
            self.G[key] = 0
        self.t = 1

    def adam(self, x, gradient_x, G, M, t, init_lr):
        """
        adam算法更新参数
        输入:
            - x:参数
            - G:梯度平方的加权移动平均
            - M:梯度的加权移动平均
            - t:迭代次数
            - init_lr:初始学习率
        """
        M = self.beta1 * M + (1 - self.beta1) * gradient_x
        G = self.beta2 * G + (1 - self.beta2) * gradient_x ** 2
        M_hat = M / (1 - self.beta1 ** t)
        G_hat = G / (1 - self.beta2 ** t)
        t += 1
        x -= init_lr / torch.sqrt(G_hat + self.epsilon) * M_hat
        return x, G, M, t

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key], self.M[key], self.t = self.adam(self.model.params[key],
                                                                                 self.model.grads[key],
                                                                                 self.G[key],
                                                                                 self.M[key],
                                                                                 self.t,
                                                                                 self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.3, 2])
model = OptimizedFunction(w)
opt = Adam(init_lr=0.2, model=model, beta1=0.9, beta2=0.99, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para5-Adam.pdf')
plt.title("Adam ")
plt.show()

        从输出结果看,Adam算法可以自适应调整学习率,参数更新更加平稳。 在增大epoch之后成功找到最优点。

1.4 完整代码 

'''
@Function: 函数x^2的不同优化算法的比较分析及2D可视化
@Author:lxy
@Date:2024/12/19
'''
from op import Op,SimpleBatchGD,Optimizer
import torch
import numpy as np
from matplotlib import pyplot as plt
import copy


# 优化函数
class OptimizedFunction(Op):
    def __init__(self, w):
        super(OptimizedFunction, self).__init__()
        self.w = w
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return torch.matmul(self.w.T, torch.tensor(torch.square(self.params['x']), dtype=torch.float32))

    def backward(self):
        self.grads['x'] = 2 * torch.multiply(self.w.T, self.params['x'])


# 训练函数,记录梯度下降过程中每轮的参数x和损失
def train_f(model, optimizer, x_init, epoch):
    """
    训练函数
    输入:
        - model:被优化函数
        - optimizer:优化器
        - x_init:x初始值
        - epoch:训练回合数
    """
    x = x_init
    all_x = []
    losses = []
    for i in range(epoch):
        all_x.append(copy.copy(x.numpy()))
        loss = model(x)
        losses.append(loss)
        model.backward()
        optimizer.step()
        x = model.params['x']
        print(all_x)
    return torch.tensor(all_x), losses


# 可视化函数,用于绘制更新轨迹
class Visualization(object):
    def __init__(self):
        """
        初始化可视化类
        """
        # 只画出参数x1和x2在区间[-5, 5]的曲线部分
        x1 = np.arange(-5, 5, 0.1)
        x2 = np.arange(-5, 5, 0.1)
        x1, x2 = np.meshgrid(x1, x2)
        self.init_x = torch.tensor([x1, x2])

    def plot_2d(self, model, x, fig_name):
        """
        可视化参数更新轨迹
        """
        fig, ax = plt.subplots(figsize=(10, 6))
        cp = ax.contourf(self.init_x[0], self.init_x[1], model(self.init_x.transpose(0, 1)),
                         colors=['#e4007f', '#f19ec2', '#e86096', '#eb7aaa', '#f6c8dc', '#f5f5f5', '#000000'])
        c = ax.contour(self.init_x[0], self.init_x[1], model(self.init_x.transpose(0, 1)), colors='black')
        cbar = fig.colorbar(cp)
        ax.plot(x[:, 0], x[:, 1], '-o', color='#000000')
        ax.plot(0, 'r*', markersize=18, color='#fefefe')

        ax.set_xlabel('$x1$')
        ax.set_ylabel('$x2$')

        ax.set_xlim((-2, 5))
        ax.set_ylim((-2, 5))
        plt.savefig(fig_name)


# 训练模型并可视化参数更新轨迹
def train_and_plot_f(model, optimizer, epoch, fig_name):
    """
    训练模型并可视化参数更新轨迹
    """
    # 设置x的初始值
    x_init = torch.tensor([3, 4], dtype=torch.float32)
    print('x1 initiate: {}, x2 initiate: {}'.format(x_init[0].numpy(), x_init[1].numpy()))
    x, losses = train_f(model, optimizer, x_init, epoch)
    print(x)
    losses = np.array(losses)

    # 展示x1、x2的更新轨迹
    vis = Visualization()
    vis.plot_2d(model, x, fig_name)

#1.SGD
# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = SimpleBatchGD(init_lr=0.2, model=model)
train_and_plot_f(model, opt, epoch=40, fig_name='opti-vis-para.pdf')
plt.title("SGD")
plt.show()
#2.AdaGrad
class Adagrad(Optimizer):
    def __init__(self, init_lr, model, epsilon):
        """
        Adagrad 优化器初始化
        输入:
            - init_lr: 初始学习率
            - model:模型,model.params存储模型参数值
            - epsilon:保持数值稳定性而设置的非常小的常数
        """
        super(Adagrad, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.epsilon = epsilon

    def adagrad(self, x, gradient_x, G, init_lr):
        """
        adagrad算法更新参数,G为参数梯度平方的累计值。
        """
        G += gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """
        参数更新
        """
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.adagrad(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = Adagrad(init_lr=0.5, model=model, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para2.pdf')
plt.title("AdaGrad")
plt.show()
#3.RMSprop
class RMSprop(Optimizer):
    def __init__(self, init_lr, model, beta, epsilon):
        """
        RMSprop优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta:衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(RMSprop, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.beta = beta
        self.epsilon = epsilon

    def rmsprop(self, x, gradient_x, G, init_lr):
        """
        rmsprop算法更新参数,G为迭代梯度平方的加权移动平均
        """
        G = self.beta * G + (1 - self.beta) * gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.rmsprop(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = RMSprop(init_lr=0.1, model=model, beta=0.9, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para3-RMSprop.pdf')
plt.title("RMSprop")
plt.show()
#4.Momentum
class Momentum(Optimizer):
    def __init__(self, init_lr, model, rho):
        """
        Momentum优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - rho:动量因子
        """
        super(Momentum, self).__init__(init_lr=init_lr, model=model)
        self.delta_x = {}
        for key in self.model.params.keys():
            self.delta_x[key] = 0
        self.rho = rho

    def momentum(self, x, gradient_x, delta_x, init_lr):
        """
        momentum算法更新参数,delta_x为梯度的加权移动平均
        """
        delta_x = self.rho * delta_x - init_lr * gradient_x
        x += delta_x
        return x, delta_x

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.delta_x[key] = self.momentum(self.model.params[key],
                                                                      self.model.grads[key],
                                                                      self.delta_x[key],
                                                                      self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.2, 2])
model = OptimizedFunction(w)
opt = Momentum(init_lr=0.1, model=model, rho=0.9)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para4-Momentum.pdf')
plt.title("Momentum")
plt.show()
#5.Adam
class Adam(Optimizer):
    def __init__(self, init_lr, model, beta1, beta2, epsilon):
        """
        Adam优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta1, beta2:移动平均的衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(Adam, self).__init__(init_lr=init_lr, model=model)
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.M, self.G = {}, {}
        for key in self.model.params.keys():
            self.M[key] = 0
            self.G[key] = 0
        self.t = 1

    def adam(self, x, gradient_x, G, M, t, init_lr):
        """
        adam算法更新参数
        输入:
            - x:参数
            - G:梯度平方的加权移动平均
            - M:梯度的加权移动平均
            - t:迭代次数
            - init_lr:初始学习率
        """
        M = self.beta1 * M + (1 - self.beta1) * gradient_x
        G = self.beta2 * G + (1 - self.beta2) * gradient_x ** 2
        M_hat = M / (1 - self.beta1 ** t)
        G_hat = G / (1 - self.beta2 ** t)
        t += 1
        x -= init_lr / torch.sqrt(G_hat + self.epsilon) * M_hat
        return x, G, M, t

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key], self.M[key], self.t = self.adam(self.model.params[key],
                                                                                 self.model.grads[key],
                                                                                 self.G[key],
                                                                                 self.M[key],
                                                                                 self.t,
                                                                                 self.init_lr)

# 固定随机种子
torch.manual_seed(0)
w = torch.tensor([0.3, 2])
model = OptimizedFunction(w)
opt = Adam(init_lr=0.2, model=model, beta1=0.9, beta2=0.99, epsilon=1e-7)
train_and_plot_f(model, opt, epoch=50, fig_name='opti-vis-para5-Adam.pdf')
plt.title("Adam ")
plt.show()

op类:

import torch
from abc import abstractmethod


class Op(object):
    def __init__(self):
        pass

    def __call__(self, inputs):
        return self.forward(inputs)

    # 输入:张量inputs
    # 输出:张量outputs
    def forward(self, inputs):
        raise NotImplementedError

    # 输入:最终输出对outputs的梯度outputs_grads
    # 输出:最终输出对inputs的梯度inputs_grads
    def backward(self, outputs_grads):
        raise NotImplementedError


# 优化器基类
class Optimizer(object):
    def __init__(self, init_lr, model):
        """
        优化器类初始化
        """
        self.init_lr = init_lr  # 初始化学习率
        self.model = model  # 模型对象

    @abstractmethod
    def step(self):
        """
        定义每次迭代如何更新参数
        """
        pass


class SimpleBatchGD(Optimizer):
    def __init__(self, init_lr, model):
        super(SimpleBatchGD, self).__init__(init_lr=init_lr, model=model)

    def step(self):
        # 参数更新
        if isinstance(self.model.params, dict):
            for key in self.model.params.keys():
                self.model.params[key] -= self.init_lr * self.model.grads[key]


class BatchGD(Optimizer):
    def __init__(self, init_lr, model):
        super(BatchGD, self).__init__(init_lr=init_lr, model=model)

    def step(self):
        # 参数更新
        for layer in self.model.layers:  # 遍历所有层
            if isinstance(layer.params, dict):
                for key in layer.params.keys():
                    layer.params[key] -= self.init_lr * layer.grads[key]


class Linear(Op):
    def __init__(self, in_features, out_features, name, weight_init=torch.randn, bias_init=torch.zeros):
        self.params = {}
        self.params['W'] = weight_init([in_features, out_features])
        self.params['b'] = bias_init([1, out_features])

        self.inputs = None
        self.grads = {}

        self.name = name

    def forward(self, inputs):
        self.inputs = inputs
        outputs = torch.matmul(self.inputs, self.params['W']) + self.params['b']
        return outputs

    def backward(self, grads):
        """
        输入:
            - grads:损失函数对当前层输出的导数
        输出:
            - 损失函数对当前层输入的导数
        """
        self.grads['W'] = torch.matmul(self.inputs.T, grads)
        self.grads['b'] = torch.sum(grads, dim=0)
        return torch.matmul(grads, self.params['W'].T)


class Logistic(Op):
    def __init__(self):
        self.inputs = None
        self.outputs = None
        self.params = None

    def forward(self, inputs):
        outputs = 1.0 / (1.0 + torch.exp(-inputs))
        self.outputs = outputs
        return outputs

    def backward(self, outputs_grads):
        # 计算Logistic激活函数对输入的导数
        outputs_grad_inputs = self.outputs * (1.0 - self.outputs)
        return outputs_grads * outputs_grad_inputs


class ReLU(Op):
    def __init__(self):
        self.inputs = None
        self.outputs = None
        self.params = None

    def forward(self, inputs):
        self.inputs = inputs
        return torch.relu(inputs)

    def backward(self, outputs_grads):
        return outputs_grads * (self.inputs > 0).float()


class MLP_3L(Op):
    def __init__(self, layers_size):
        self.fc1 = Linear(layers_size[0], layers_size[1], name='fc1')
        # ReLU激活函数
        self.act_fn1 = ReLU()
        self.fc2 = Linear(layers_size[1], layers_size[2], name='fc2')
        self.act_fn2 = ReLU()
        self.fc3 = Linear(layers_size[2], layers_size[3], name='fc3')
        self.layers = [self.fc1, self.act_fn1, self.fc2, self.act_fn2, self.fc3]

    def __call__(self, X):
        return self.forward(X)

    def forward(self, X):
        z1 = self.fc1(X)
        a1 = self.act_fn1(z1)
        z2 = self.fc2(a1)
        a2 = self.act_fn2(z2)
        z3 = self.fc3(a2)
        return z3

    def backward(self, loss_grad_z3):
        loss_grad_a2 = self.fc3.backward(loss_grad_z3)
        loss_grad_z2 = self.act_fn2.backward(loss_grad_a2)
        loss_grad_a1 = self.fc2.backward(loss_grad_z2)
        loss_grad_z1 = self.act_fn1.backward(loss_grad_a1)
        loss_grad_inputs = self.fc1.backward(loss_grad_z1)
        return loss_grad_inputs

1.5 函数 \frac{1}{20}x^2+y^2的优化算法比较

'''
@Function: 函数x^2/20+y^2的不同优化算法的比较分析及2D可视化
@Author:lxy
@Date:2024/12/19
'''
import numpy as np
import matplotlib.pyplot as plt
from collections import OrderedDict

class SGD:
    """随机梯度下降法(Stochastic Gradient Descent)"""

    def __init__(self, lr=0.01):
        self.lr = lr

    def update(self, params, grads):
        for key in params.keys():
            params[key] -= self.lr * grads[key]


class Momentum:
    """Momentum SGD"""

    def __init__(self, lr=0.01, momentum=0.9):
        self.lr = lr
        self.momentum = momentum
        self.v = None

    def update(self, params, grads):
        if self.v is None:
            self.v = {}
            for key, val in params.items():
                self.v[key] = np.zeros_like(val)

        for key in params.keys():
            self.v[key] = self.momentum * self.v[key] - self.lr * grads[key]
            params[key] += self.v[key]


class Nesterov:
    """Nesterov's Accelerated Gradient (http://arxiv.org/abs/1212.0901)"""

    def __init__(self, lr=0.01, momentum=0.9):
        self.lr = lr
        self.momentum = momentum
        self.v = None

    def update(self, params, grads):
        if self.v is None:
            self.v = {}
            for key, val in params.items():
                self.v[key] = np.zeros_like(val)

        for key in params.keys():
            self.v[key] *= self.momentum
            self.v[key] -= self.lr * grads[key]
            params[key] += self.momentum * self.momentum * self.v[key]
            params[key] -= (1 + self.momentum) * self.lr * grads[key]


class AdaGrad:
    """AdaGrad"""

    def __init__(self, lr=0.01):
        self.lr = lr
        self.h = None

    def update(self, params, grads):
        if self.h is None:
            self.h = {}
            for key, val in params.items():
                self.h[key] = np.zeros_like(val)

        for key in params.keys():
            self.h[key] += grads[key] * grads[key]
            params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)


class RMSprop:
    """RMSprop"""

    def __init__(self, lr=0.01, decay_rate=0.99):
        self.lr = lr
        self.decay_rate = decay_rate
        self.h = None

    def update(self, params, grads):
        if self.h is None:
            self.h = {}
            for key, val in params.items():
                self.h[key] = np.zeros_like(val)

        for key in params.keys():
            self.h[key] *= self.decay_rate
            self.h[key] += (1 - self.decay_rate) * grads[key] * grads[key]
            params[key] -= self.lr * grads[key] / (np.sqrt(self.h[key]) + 1e-7)


class Adam:
    """Adam (http://arxiv.org/abs/1412.6980v8)"""

    def __init__(self, lr=0.001, beta1=0.9, beta2=0.999):
        self.lr = lr
        self.beta1 = beta1
        self.beta2 = beta2
        self.iter = 0
        self.m = None
        self.v = None

    def update(self, params, grads):
        if self.m is None:
            self.m, self.v = {}, {}
            for key, val in params.items():
                self.m[key] = np.zeros_like(val)
                self.v[key] = np.zeros_like(val)

        self.iter += 1
        lr_t = self.lr * np.sqrt(1.0 - self.beta2 ** self.iter) / (1.0 - self.beta1 ** self.iter)

        for key in params.keys():
            self.m[key] += (1 - self.beta1) * (grads[key] - self.m[key])
            self.v[key] += (1 - self.beta2) * (grads[key] ** 2 - self.v[key])

            params[key] -= lr_t * self.m[key] / (np.sqrt(self.v[key]) + 1e-7)

def f(x, y):
    return x ** 2 / 20.0 + y ** 2

def df(x, y):
    return x / 10.0, 2.0 * y

init_pos = (-7.0, 2.0)
params = {}
params['x'], params['y'] = init_pos[0], init_pos[1]
grads = {}
grads['x'], grads['y'] = 0, 0

learningrate = [0.9, 0.3, 0.3, 0.6, 0.6, 0.6, 0.6]
optimizers = OrderedDict() #一个有序的字典
optimizers["SGD"] = SGD(lr=learningrate[0])
optimizers["Momentum"] = Momentum(lr=learningrate[1])
optimizers["Nesterov"] = Nesterov(lr=learningrate[2])
optimizers["AdaGrad"] = AdaGrad(lr=learningrate[3])
optimizers["RMSprop"] = RMSprop(lr=learningrate[4])
optimizers["Adam"] = Adam(lr=learningrate[5])

idx = 1
id_lr = 0

for key in optimizers:
    optimizer = optimizers[key]
    lr = learningrate[id_lr]
    id_lr = id_lr + 1
    x_history = []
    y_history = []
    params['x'], params['y'] = init_pos[0], init_pos[1]

    for i in range(30):
        x_history.append(params['x'])
        y_history.append(params['y'])

        grads['x'], grads['y'] = df(params['x'], params['y'])
        optimizer.update(params, grads)

    x = np.arange(-10, 10, 0.01)
    y = np.arange(-5, 5, 0.01)

    X, Y = np.meshgrid(x, y)
    Z = f(X, Y)
    # for simple contour line
    mask = Z > 7
    Z[mask] = 0   #这个操作可以使高度大于 7 的点在绘制等高线时被忽略掉,只画出高度小于等于 7 的部分,从而得到一个简单的等高线图。

    # plot
    plt.subplot(2, 3, idx)
    idx += 1
    plt.plot(x_history, y_history, 'o-',markersize=4, color="r")
    # plt.contour(X, Y, Z)  # 绘制等高线
    plt.contour(X, Y, Z, cmap='rainbow')  # 颜色填充
    plt.ylim(-10, 10)
    plt.xlim(-10, 10)
    plt.plot(0, 0, '+')
    # plt.axis('off')
    # plt.title(key+'\nlr='+str(lr), fontstyle='italic')
    plt.text(0, 10, key + '\nlr=' + str(lr), fontsize=10, color="b",
             verticalalignment='top', horizontalalignment='center', fontstyle='italic')
    plt.xlabel("x")
    plt.ylabel("y")

plt.subplots_adjust(wspace=0, hspace=0)  # 调整子图间距
plt.show()

RMSprop的效果比较好,迭代了几次就到终点了。

2 不同优化算法比较分析-3D可视化实验

2.1 函数3D可视化

'''
@Function:两个函数的3D可视化
@Author;lxy
@Date:2024/12/19
'''
import numpy as np
from matplotlib import pyplot as plt
import torch
from op import Op

#1.函数3D可视化
#第一个函数
class OptimizedFunction3D1(Op):
    def __init__(self):
        super(OptimizedFunction3D1, self).__init__()
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return x[0] ** 2 + x[1] ** 2 + x[1] ** 3 + x[0] * x[1]

    def backward(self):
        x = self.params['x']
        gradient1 = 2 * x[0] + x[1]
        gradient2 = 2 * x[1] + 3 * x[1] ** 2 + x[0]
        grad1 = torch.Tensor([gradient1])
        grad2 = torch.Tensor([gradient2])
        self.grads['x'] = torch.cat([grad1, grad2])

#第二个函数
class OptimizedFunction3D2(Op):
    def __init__(self):
        super(OptimizedFunction3D2, self).__init__()
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return x[0] * x[0] / 20 + x[1] * x[1] / 1  # x[0] ** 2 + x[1] ** 2 + x[1] ** 3 + x[0] * x[1]

    def backward(self):
        x = self.params['x']
        gradient1 = 2 * x[0] / 20
        gradient2 = 2 * x[1] / 1
        grad1 = torch.Tensor([gradient1])
        grad2 = torch.Tensor([gradient2])
        self.grads['x'] = torch.cat([grad1, grad2])


# 使用numpy.meshgrid生成x1,x2矩阵,矩阵的每一行为[-3, 3],以0.1为间隔的数值
x1 = np.arange(-3, 3, 0.1)
x2 = np.arange(-3, 3, 0.1)
x1, x2 = np.meshgrid(x1, x2)#网格
init_x = torch.Tensor(np.array([x1, x2]))

model1 = OptimizedFunction3D1()#实例化
model2 = OptimizedFunction3D2()#实例化

fig = plt.figure()
#绘制第一个函数的3D图像
ax1 = fig.add_subplot(121, projection='3d')
y1 = model1.forward(init_x)
ax1.plot_surface(x1, x2, y1.detach(), cmap='rainbow')
ax1.set_xlabel('x1')
ax1.set_ylabel('x2')
ax1.set_zlabel('z1')
ax1.set_title('Optimized Function1 3D')

#绘制第二个函数的3D图像
ax2 = fig.add_subplot(122, projection='3d')
y2 = model2.forward(init_x)
ax2.plot_surface(x1, x2, y2.detach(), cmap='coolwarm')
ax2.set_xlabel('x1')
ax2.set_ylabel('x2')
ax2.set_zlabel('z2')
ax2.set_title('Optimized Function2 3D')

plt.show()

  x[0]^{2}+x[1]^2+x[1]^3+x[0]*x[1]

  \frac{1}{20}x^2+y^2

2.2 加入不同优化算法,画出轨迹

        在 2D可视化实验的设定基础上添加3D绘图的要素,主要是构建 Visualization3D 类继承自 animation.FuncAnimation,初始化接收轨迹数据等并创建绘图对象,init_animation 方法初始化绘图数据,animate 方法更新绘图数据实现动态绘制。

2.2.1  x[0]^{2}+x[1]^2+x[1]^3+x[0]*x[1]

'''
@Function:第一个函数的不同优化算法的3D可视化
@Author;lxy
@Date:2024/12/19
'''
#2.加入优化算法,画出轨迹
import torch
import numpy as np
import copy
from matplotlib import pyplot as plt
from matplotlib import animation
from itertools import zip_longest


class Op(object):
    def __init__(self):
        pass

    def __call__(self, inputs):
        return self.forward(inputs)

    # 输入:张量inputs
    # 输出:张量outputs
    def forward(self, inputs):
        # return outputs
        raise NotImplementedError

    # 输入:最终输出对outputs的梯度outputs_grads
    # 输出:最终输出对inputs的梯度inputs_grads
    def backward(self, outputs_grads):
        # return inputs_grads
        raise NotImplementedError


class Optimizer(object):  # 优化器基类
    def __init__(self, init_lr, model):
        """
        优化器类初始化
        """
        # 初始化学习率,用于参数更新的计算
        self.init_lr = init_lr
        # 指定优化器需要优化的模型
        self.model = model

    def step(self):
        """
        定义每次迭代如何更新参数
        """
        pass


class SimpleBatchGD(Optimizer):
    def __init__(self, init_lr, model):
        super(SimpleBatchGD, self).__init__(init_lr=init_lr, model=model)

    def step(self):
        # 参数更新
        if isinstance(self.model.params, dict):
            for key in self.model.params.keys():
                self.model.params[key] = self.model.params[key] - self.init_lr * self.model.grads[key]


class Adagrad(Optimizer):
    def __init__(self, init_lr, model, epsilon):
        """
        Adagrad 优化器初始化
        输入:
            - init_lr: 初始学习率 - model:模型,model.params存储模型参数值  - epsilon:保持数值稳定性而设置的非常小的常数
        """
        super(Adagrad, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.epsilon = epsilon

    def adagrad(self, x, gradient_x, G, init_lr):
        """
        adagrad算法更新参数,G为参数梯度平方的累计值。
        """
        G += gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """
        参数更新
        """
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.adagrad(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class RMSprop(Optimizer):
    def __init__(self, init_lr, model, beta, epsilon):
        """
        RMSprop优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta:衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(RMSprop, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.beta = beta
        self.epsilon = epsilon

    def rmsprop(self, x, gradient_x, G, init_lr):
        """
        rmsprop算法更新参数,G为迭代梯度平方的加权移动平均
        """
        G = self.beta * G + (1 - self.beta) * gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.rmsprop(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class Momentum(Optimizer):
    def __init__(self, init_lr, model, rho):
        """
        Momentum优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - rho:动量因子
        """
        super(Momentum, self).__init__(init_lr=init_lr, model=model)
        self.delta_x = {}
        for key in self.model.params.keys():
            self.delta_x[key] = 0
        self.rho = rho

    def momentum(self, x, gradient_x, delta_x, init_lr):
        """
        momentum算法更新参数,delta_x为梯度的加权移动平均
        """
        delta_x = self.rho * delta_x - init_lr * gradient_x
        x += delta_x
        return x, delta_x

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.delta_x[key] = self.momentum(self.model.params[key],
                                                                      self.model.grads[key],
                                                                      self.delta_x[key],
                                                                      self.init_lr)


class Adam(Optimizer):
    def __init__(self, init_lr, model, beta1, beta2, epsilon):
        """
        Adam优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta1, beta2:移动平均的衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(Adam, self).__init__(init_lr=init_lr, model=model)
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.M, self.G = {}, {}
        for key in self.model.params.keys():
            self.M[key] = 0
            self.G[key] = 0
        self.t = 1

    def adam(self, x, gradient_x, G, M, t, init_lr):
        """
        adam算法更新参数
        输入:
            - x:参数
            - G:梯度平方的加权移动平均
            - M:梯度的加权移动平均
            - t:迭代次数
            - init_lr:初始学习率
        """
        M = self.beta1 * M + (1 - self.beta1) * gradient_x
        G = self.beta2 * G + (1 - self.beta2) * gradient_x ** 2
        M_hat = M / (1 - self.beta1 ** t)
        G_hat = G / (1 - self.beta2 ** t)
        t += 1
        x -= init_lr / torch.sqrt(G_hat + self.epsilon) * M_hat
        return x, G, M, t

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key], self.M[key], self.t = self.adam(self.model.params[key],
                                                                                 self.model.grads[key],
                                                                                 self.G[key],
                                                                                 self.M[key],
                                                                                 self.t,
                                                                                 self.init_lr)

class Visualization3D(animation.FuncAnimation):
    """    绘制动态图像,可视化参数更新轨迹    """

    def __init__(self, *xy_values, z_values, labels=[], colors=[], fig, ax, interval=600, blit=True, **kwargs):
        """
        初始化3d可视化类
        输入:
            xy_values:三维中x,y维度的值
            z_values:三维中z维度的值
            labels:每个参数更新轨迹的标签
            colors:每个轨迹的颜色
            interval:帧之间的延迟(以毫秒为单位)
            blit:是否优化绘图
        """
        self.fig = fig
        self.ax = ax
        self.xy_values = xy_values
        self.z_values = z_values

        frames = max(xy_value.shape[0] for xy_value in xy_values)
        self.lines = [ax.plot([], [], [], label=label, color=color, lw=2)[0]
                      for _, label, color in zip_longest(xy_values, labels, colors)]
        super(Visualization3D, self).__init__(fig, self.animate, init_func=self.init_animation, frames=frames,
                                              interval=interval, blit=blit, **kwargs)

    def init_animation(self):
        # 数值初始化
        for line in self.lines:
            line.set_data([], [])
            # line.set_3d_properties(np.asarray([]))  # 源程序中有这一行,加上会报错。 Edit by David 2022.12.4
        return self.lines

    def animate(self, i):
        # 将x,y,z三个数据传入,绘制三维图像
        for line, xy_value, z_value in zip(self.lines, self.xy_values, self.z_values):
            line.set_data(xy_value[:i, 0], xy_value[:i, 1])
            line.set_3d_properties(z_value[:i])

        return self.lines

class OptimizedFunction3D(Op):
    def __init__(self):
        super(OptimizedFunction3D, self).__init__()
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return x[0] ** 2 + x[1] ** 2 + x[1] ** 3 + x[0] * x[1]

    def backward(self):
        x = self.params['x']
        gradient1 = 2 * x[0] + x[1]
        gradient2 = 2 * x[1] + 3 * x[1] ** 2 + x[0]
        grad1 = torch.Tensor([gradient1])
        grad2 = torch.Tensor([gradient2])
        self.grads['x'] = torch.cat([grad1, grad2])


def train_f(model, optimizer, x_init, epoch):
    x = x_init
    all_x = []
    losses = []
    for i in range(epoch):
        all_x.append(copy.deepcopy(x.numpy()))  # 浅拷贝 改为 深拷贝, 否则List的原值会被改变。 Edit by David 2022.12.4.
        loss = model(x)
        losses.append(loss)
        model.backward()
        optimizer.step()
        x = model.params['x']
    return torch.Tensor(np.array(all_x)), losses

SGD_lr=0.05
Adagrad_lr=0.05
RMSprop_lr=0.05
Momentum_lr=0.01
Adam_lr=0.05

# 构建5个模型,分别配备不同的优化器
model1 = OptimizedFunction3D()
opt_gd = SimpleBatchGD(init_lr=SGD_lr, model=model1)

model2 = OptimizedFunction3D()
opt_adagrad = Adagrad(init_lr=Adagrad_lr, model=model2, epsilon=1e-7)

model3 = OptimizedFunction3D()
opt_rmsprop = RMSprop(init_lr=RMSprop_lr, model=model3, beta=0.9, epsilon=1e-7)

model4 = OptimizedFunction3D()
opt_momentum = Momentum(init_lr=Momentum_lr, model=model4, rho=0.9)

model5 = OptimizedFunction3D()
opt_adam = Adam(init_lr=Adam_lr, model=model5, beta1=0.9, beta2=0.99, epsilon=1e-7)

models = [model1, model2, model3, model4, model5]
opts = [opt_gd, opt_adagrad, opt_rmsprop, opt_momentum, opt_adam]

x_all_opts = []
z_all_opts = []

# 使用不同优化器训练
for model, opt in zip(models, opts):
    x_init = torch.FloatTensor([2, 3])
    x_one_opt, z_one_opt = train_f(model, opt, x_init, 150)  # epoch
    # 保存参数值
    x_all_opts.append(x_one_opt.numpy())
    z_all_opts.append(np.squeeze(z_one_opt))

# 使用numpy.meshgrid生成x1,x2矩阵,矩阵的每一行为[-3, 3],以0.1为间隔的数值
x1 = np.arange(-3, 3, 0.1)
x2 = np.arange(-3, 3, 0.1)
x1, x2 = np.meshgrid(x1, x2)
init_x = torch.Tensor(np.array([x1, x2]))

model = OptimizedFunction3D()

# 绘制 f_3d函数 的 三维图像
fig = plt.figure()
ax = plt.axes(projection='3d')
X = init_x[0].numpy()
Y = init_x[1].numpy()
Z = model(init_x).numpy()  # 改为 model(init_x).numpy() David 2022.12.4
ax.plot_surface(X, Y, Z, cmap='rainbow')

ax.set_xlabel('x1')
ax.set_ylabel('x2')
ax.set_zlabel('f(x1,x2)')

labels = ['SGD lr='+str(SGD_lr), 'AdaGrad lr='+str(Adagrad_lr), 'RMSprop lr='+str(RMSprop_lr), 'Momentum lr='+str(Momentum_lr), 'Adam lr='+str(Adam_lr)]
colors = ['#f6373c', '#f6f237', '#45f637', '#37f0f6', '#000000']

animator = Visualization3D(*x_all_opts, z_values=z_all_opts, labels=labels, colors=colors, fig=fig, ax=ax)
ax.legend(loc='upper left')

plt.show()

animator.save('animation.gif')  # 效果不好,估计被挡住了…… 有待进一步提高 Edit by David 2022.12.4

学习率相对较小: 

SGD_lr=0.05
Adagrad_lr=0.05
RMSprop_lr=0.05
Momentum_lr=0.01
Adam_lr=0.05

 

效果都不是很好,适当学习率增大:

SGD_lr=0.3
Adagrad_lr=0.3
RMSprop_lr=0.3
Momentum_lr=0.01
Adam_lr=0.3

 

【分析】

        针对这个函数,在相同的学习率时,

  • SGD的更新方向比较稳定,速度很快,直接跳过了局部最小值。
  • AdaGrad更新速度很慢,而且越往后由于学习率衰减,它几乎不更新了。
  •  RMSprop表现得不是很好,下降速度慢,而且陷入了函数中比较平缓的部分,最后在局部最小值处震荡。
  • Momentum下降是速度很快,而且方向也比较稳定,表现很好
  • Adam刚开始速度较慢,后来就变快了 ,并跳过了局部极小值处。

2.2.2  \frac{1}{20}x^2+y^2

'''
@Function:第二个函数的不同优化算法的3D可视化
@Author;lxy
@Date:2024/12/19
'''
import torch
import numpy as np
import copy
from matplotlib import pyplot as plt
from matplotlib import animation
from itertools import zip_longest
from matplotlib import cm


class Op(object):
    def __init__(self):
        pass

    def __call__(self, inputs):
        return self.forward(inputs)

    # 输入:张量inputs
    # 输出:张量outputs
    def forward(self, inputs):
        # return outputs
        raise NotImplementedError

    # 输入:最终输出对outputs的梯度outputs_grads
    # 输出:最终输出对inputs的梯度inputs_grads
    def backward(self, outputs_grads):
        # return inputs_grads
        raise NotImplementedError


class Optimizer(object):  # 优化器基类
    def __init__(self, init_lr, model):
        """
        优化器类初始化
        """
        # 初始化学习率,用于参数更新的计算
        self.init_lr = init_lr
        # 指定优化器需要优化的模型
        self.model = model

    def step(self):
        """
        定义每次迭代如何更新参数
        """
        pass


class SimpleBatchGD(Optimizer):
    def __init__(self, init_lr, model):
        super(SimpleBatchGD, self).__init__(init_lr=init_lr, model=model)

    def step(self):
        # 参数更新
        if isinstance(self.model.params, dict):
            for key in self.model.params.keys():
                self.model.params[key] = self.model.params[key] - self.init_lr * self.model.grads[key]


class Adagrad(Optimizer):
    def __init__(self, init_lr, model, epsilon):
        """
        Adagrad 优化器初始化
        输入:
            - init_lr: 初始学习率 - model:模型,model.params存储模型参数值  - epsilon:保持数值稳定性而设置的非常小的常数
        """
        super(Adagrad, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.epsilon = epsilon

    def adagrad(self, x, gradient_x, G, init_lr):
        """
        adagrad算法更新参数,G为参数梯度平方的累计值。
        """
        G += gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """
        参数更新
        """
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.adagrad(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class RMSprop(Optimizer):
    def __init__(self, init_lr, model, beta, epsilon):
        """
        RMSprop优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta:衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(RMSprop, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.beta = beta
        self.epsilon = epsilon

    def rmsprop(self, x, gradient_x, G, init_lr):
        """
        rmsprop算法更新参数,G为迭代梯度平方的加权移动平均
        """
        G = self.beta * G + (1 - self.beta) * gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.rmsprop(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class Momentum(Optimizer):
    def __init__(self, init_lr, model, rho):
        """
        Momentum优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - rho:动量因子
        """
        super(Momentum, self).__init__(init_lr=init_lr, model=model)
        self.delta_x = {}
        for key in self.model.params.keys():
            self.delta_x[key] = 0
        self.rho = rho

    def momentum(self, x, gradient_x, delta_x, init_lr):
        """
        momentum算法更新参数,delta_x为梯度的加权移动平均
        """
        delta_x = self.rho * delta_x - init_lr * gradient_x
        x += delta_x
        return x, delta_x

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.delta_x[key] = self.momentum(self.model.params[key],
                                                                      self.model.grads[key],
                                                                      self.delta_x[key],
                                                                      self.init_lr)


class Adam(Optimizer):
    def __init__(self, init_lr, model, beta1, beta2, epsilon):
        """
        Adam优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta1, beta2:移动平均的衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(Adam, self).__init__(init_lr=init_lr, model=model)
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.M, self.G = {}, {}
        for key in self.model.params.keys():
            self.M[key] = 0
            self.G[key] = 0
        self.t = 1

    def adam(self, x, gradient_x, G, M, t, init_lr):
        """
        adam算法更新参数
        输入:
            - x:参数
            - G:梯度平方的加权移动平均
            - M:梯度的加权移动平均
            - t:迭代次数
            - init_lr:初始学习率
        """
        M = self.beta1 * M + (1 - self.beta1) * gradient_x
        G = self.beta2 * G + (1 - self.beta2) * gradient_x ** 2
        M_hat = M / (1 - self.beta1 ** t)
        G_hat = G / (1 - self.beta2 ** t)
        t += 1
        x -= init_lr / torch.sqrt(G_hat + self.epsilon) * M_hat
        return x, G, M, t

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key], self.M[key], self.t = self.adam(self.model.params[key],
                                                                                 self.model.grads[key],
                                                                                 self.G[key],
                                                                                 self.M[key],
                                                                                 self.t,
                                                                                 self.init_lr)


class OptimizedFunction3D(Op):
    def __init__(self):
        super(OptimizedFunction3D, self).__init__()
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return x[0] * x[0] / 20 + x[1] * x[1] / 1  # x[0] ** 2 + x[1] ** 2 + x[1] ** 3 + x[0] * x[1]

    def backward(self):
        x = self.params['x']
        gradient1 = 2 * x[0] / 20
        gradient2 = 2 * x[1] / 1
        grad1 = torch.Tensor([gradient1])
        grad2 = torch.Tensor([gradient2])
        self.grads['x'] = torch.cat([grad1, grad2])


class Visualization3D(animation.FuncAnimation):
    """    绘制动态图像,可视化参数更新轨迹    """

    def __init__(self, *xy_values, z_values, labels=[], colors=[], fig, ax, interval=100, blit=True, **kwargs):
        """
        初始化3d可视化类
        输入:
            xy_values:三维中x,y维度的值
            z_values:三维中z维度的值
            labels:每个参数更新轨迹的标签
            colors:每个轨迹的颜色
            interval:帧之间的延迟(以毫秒为单位)
            blit:是否优化绘图
        """
        self.fig = fig
        self.ax = ax
        self.xy_values = xy_values
        self.z_values = z_values

        frames = max(xy_value.shape[0] for xy_value in xy_values)

        self.lines = [ax.plot([], [], [], label=label, color=color, lw=2)[0]
                      for _, label, color in zip_longest(xy_values, labels, colors)]
        self.points = [ax.plot([], [], [], color=color, markeredgewidth=1, markeredgecolor='black', marker='o')[0]
                       for _, color in zip_longest(xy_values, colors)]
        # print(self.lines)
        super(Visualization3D, self).__init__(fig, self.animate, init_func=self.init_animation, frames=frames,
                                              interval=interval, blit=blit, **kwargs)

    def init_animation(self):
        # 数值初始化
        for line in self.lines:
            line.set_data_3d([], [], [])
        for point in self.points:
            point.set_data_3d([], [], [])
        return self.points + self.lines

    def animate(self, i):
        # 将x,y,z三个数据传入,绘制三维图像
        for line, xy_value, z_value in zip(self.lines, self.xy_values, self.z_values):
            line.set_data_3d(xy_value[:i, 0], xy_value[:i, 1], z_value[:i])
        for point, xy_value, z_value in zip(self.points, self.xy_values, self.z_values):
           #point.set_data_3d(xy_value[i, 0], xy_value[i, 1], z_value[i])
            point.set_data_3d([xy_value[i, 0]], [xy_value[i, 1]], [z_value[i]])

        return self.points + self.lines


def train_f(model, optimizer, x_init, epoch):
    x = x_init
    all_x = []
    losses = []
    for i in range(epoch):
        all_x.append(copy.deepcopy(x.numpy()))  # 浅拷贝 改为 深拷贝, 否则List的原值会被改变。 Edit by David 2022.12.4.
        loss = model(x)
        losses.append(loss)
        model.backward()
        optimizer.step()
        x = model.params['x']
    return torch.Tensor(np.array(all_x)), losses


# 构建5个模型,分别配备不同的优化器
model1 = OptimizedFunction3D()
opt_gd = SimpleBatchGD(init_lr=0.95, model=model1)

model2 = OptimizedFunction3D()
opt_adagrad = Adagrad(init_lr=1.5, model=model2, epsilon=1e-7)

model3 = OptimizedFunction3D()
opt_rmsprop = RMSprop(init_lr=0.05, model=model3, beta=0.9, epsilon=1e-7)

model4 = OptimizedFunction3D()
opt_momentum = Momentum(init_lr=0.1, model=model4, rho=0.9)

model5 = OptimizedFunction3D()
opt_adam = Adam(init_lr=0.3, model=model5, beta1=0.9, beta2=0.99, epsilon=1e-7)

models = [model1, model2, model3, model4, model5]
opts = [opt_gd, opt_adagrad, opt_rmsprop, opt_momentum, opt_adam]

x_all_opts = []
z_all_opts = []

# 使用不同优化器训练

for model, opt in zip(models, opts):
    x_init = torch.FloatTensor([-7, 2])
    x_one_opt, z_one_opt = train_f(model, opt, x_init, 100)  # epoch
    # 保存参数值
    x_all_opts.append(x_one_opt.numpy())
    z_all_opts.append(np.squeeze(z_one_opt))

# 使用numpy.meshgrid生成x1,x2矩阵,矩阵的每一行为[-3, 3],以0.1为间隔的数值
x1 = np.arange(-10, 10, 0.01)
x2 = np.arange(-5, 5, 0.01)
x1, x2 = np.meshgrid(x1, x2)
init_x = torch.Tensor(np.array([x1, x2]))

model = OptimizedFunction3D()

# 绘制 f_3d函数 的 三维图像
fig = plt.figure()
ax = plt.axes(projection='3d')
X = init_x[0].numpy()
Y = init_x[1].numpy()
Z = model(init_x).numpy()  # 改为 model(init_x).numpy() David 2022.12.4
surf = ax.plot_surface(X, Y, Z, edgecolor='grey', cmap=cm.coolwarm)
# fig.colorbar(surf, shrink=0.5, aspect=1)
# ax.set_zlim(-3, 2)
ax.set_xlabel('x1')
ax.set_ylabel('x2')
ax.set_zlabel('f(x1,x2)')

labels = ['SGD', 'AdaGrad', 'RMSprop', 'Momentum', 'Adam']
colors = ['#8B0000', '#0000FF', '#000000', '#008B00', '#FF0000']

animator = Visualization3D(*x_all_opts, z_values=z_all_opts, labels=labels, colors=colors, fig=fig, ax=ax)
ax.legend(loc='upper right')

plt.show()
# animator.save('teaser' + '.gif', writer='imagemagick',fps=10) # 效果不好,估计被挡住了…… 有待进一步提高 Edit by David 2022.12.4
# save不好用,不费劲了,安装个软件做gif https://pc.qq.com/detail/13/detail_23913.html

【分析】

SGD:路线呈“之”字形下降,刚开始下降很快,在接近最优解时,轨迹会越来越平缓,收敛速度相对较慢。
Adagrad:随着训练的进行不断调整学习率,轨迹比梯度下降法和动量法更加平滑,且收敛速度相对较快,能够对不同参数的不同变化范围进行适当的调整,使得算法更容易收敛到全局最优解。
RMSprop:运行轨迹与Adagrad的差不多,比较平稳。
Momentum:在更新方向上增加了惯性,跳出局部极值,它的轨迹比梯度下降法更加平滑,在接近最优解时收敛速度也较快。
Adam:结合了动量法和RMSprop,运行速度较快,轨迹也比较平滑。

3 复现CS231经典动画

'''
@Function:cs231函数的不同优化算法的3D可视化
@Author;lxy
@Date:2024/12/19
'''
import torch
import numpy as np
import copy
from matplotlib import pyplot as plt
from matplotlib import animation
from itertools import zip_longest
from matplotlib import cm


class Op(object):
    def __init__(self):
        pass

    def __call__(self, inputs):
        return self.forward(inputs)

    # 输入:张量inputs
    # 输出:张量outputs
    def forward(self, inputs):
        # return outputs
        raise NotImplementedError

    # 输入:最终输出对outputs的梯度outputs_grads
    # 输出:最终输出对inputs的梯度inputs_grads
    def backward(self, outputs_grads):
        # return inputs_grads
        raise NotImplementedError


class Optimizer(object):  # 优化器基类
    def __init__(self, init_lr, model):
        """
        优化器类初始化
        """
        # 初始化学习率,用于参数更新的计算
        self.init_lr = init_lr
        # 指定优化器需要优化的模型
        self.model = model

    def step(self):
        """
        定义每次迭代如何更新参数
        """
        pass


class SimpleBatchGD(Optimizer):
    def __init__(self, init_lr, model):
        super(SimpleBatchGD, self).__init__(init_lr=init_lr, model=model)

    def step(self):
        # 参数更新
        if isinstance(self.model.params, dict):
            for key in self.model.params.keys():
                self.model.params[key] = self.model.params[key] - self.init_lr * self.model.grads[key]


class Adagrad(Optimizer):
    def __init__(self, init_lr, model, epsilon):
        """
        Adagrad 优化器初始化
        输入:
            - init_lr: 初始学习率 - model:模型,model.params存储模型参数值  - epsilon:保持数值稳定性而设置的非常小的常数
        """
        super(Adagrad, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.epsilon = epsilon

    def adagrad(self, x, gradient_x, G, init_lr):
        """
        adagrad算法更新参数,G为参数梯度平方的累计值。
        """
        G += gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """
        参数更新
        """
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.adagrad(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class RMSprop(Optimizer):
    def __init__(self, init_lr, model, beta, epsilon):
        """
        RMSprop优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta:衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(RMSprop, self).__init__(init_lr=init_lr, model=model)
        self.G = {}
        for key in self.model.params.keys():
            self.G[key] = 0
        self.beta = beta
        self.epsilon = epsilon

    def rmsprop(self, x, gradient_x, G, init_lr):
        """
        rmsprop算法更新参数,G为迭代梯度平方的加权移动平均
        """
        G = self.beta * G + (1 - self.beta) * gradient_x ** 2
        x -= init_lr / torch.sqrt(G + self.epsilon) * gradient_x
        return x, G

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key] = self.rmsprop(self.model.params[key],
                                                               self.model.grads[key],
                                                               self.G[key],
                                                               self.init_lr)


class Momentum(Optimizer):
    def __init__(self, init_lr, model, rho):
        """
        Momentum优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - rho:动量因子
        """
        super(Momentum, self).__init__(init_lr=init_lr, model=model)
        self.delta_x = {}
        for key in self.model.params.keys():
            self.delta_x[key] = 0
        self.rho = rho

    def momentum(self, x, gradient_x, delta_x, init_lr):
        """
        momentum算法更新参数,delta_x为梯度的加权移动平均
        """
        delta_x = self.rho * delta_x - init_lr * gradient_x
        x += delta_x
        return x, delta_x

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.delta_x[key] = self.momentum(self.model.params[key],
                                                                      self.model.grads[key],
                                                                      self.delta_x[key],
                                                                      self.init_lr)


class Adam(Optimizer):
    def __init__(self, init_lr, model, beta1, beta2, epsilon):
        """
        Adam优化器初始化
        输入:
            - init_lr:初始学习率
            - model:模型,model.params存储模型参数值
            - beta1, beta2:移动平均的衰减率
            - epsilon:保持数值稳定性而设置的常数
        """
        super(Adam, self).__init__(init_lr=init_lr, model=model)
        self.beta1 = beta1
        self.beta2 = beta2
        self.epsilon = epsilon
        self.M, self.G = {}, {}
        for key in self.model.params.keys():
            self.M[key] = 0
            self.G[key] = 0
        self.t = 1

    def adam(self, x, gradient_x, G, M, t, init_lr):
        """
        adam算法更新参数
        输入:
            - x:参数
            - G:梯度平方的加权移动平均
            - M:梯度的加权移动平均
            - t:迭代次数
            - init_lr:初始学习率
        """
        M = self.beta1 * M + (1 - self.beta1) * gradient_x
        G = self.beta2 * G + (1 - self.beta2) * gradient_x ** 2
        M_hat = M / (1 - self.beta1 ** t)
        G_hat = G / (1 - self.beta2 ** t)
        t += 1
        x -= init_lr / torch.sqrt(G_hat + self.epsilon) * M_hat
        return x, G, M, t

    def step(self):
        """参数更新"""
        for key in self.model.params.keys():
            self.model.params[key], self.G[key], self.M[key], self.t = self.adam(self.model.params[key],
                                                                                 self.model.grads[key],
                                                                                 self.G[key],
                                                                                 self.M[key],
                                                                                 self.t,
                                                                                 self.init_lr)


class OptimizedFunction3D(Op):
    def __init__(self):
        super(OptimizedFunction3D, self).__init__()
        self.params = {'x': 0}
        self.grads = {'x': 0}

    def forward(self, x):
        self.params['x'] = x
        return - x[0] * x[0] / 2 + x[1] * x[1] / 1  # x[0] ** 2 + x[1] ** 2 + x[1] ** 3 + x[0] * x[1]

    def backward(self):
        x = self.params['x']
        gradient1 = - 2 * x[0] / 2
        gradient2 = 2 * x[1] / 1
        grad1 = torch.Tensor([gradient1])
        grad2 = torch.Tensor([gradient2])
        self.grads['x'] = torch.cat([grad1, grad2])


class Visualization3D(animation.FuncAnimation):
    """    绘制动态图像,可视化参数更新轨迹    """

    def __init__(self, *xy_values, z_values, labels=[], colors=[], fig, ax, interval=100, blit=True, **kwargs):
        """
        初始化3d可视化类
        输入:
            xy_values:三维中x,y维度的值
            z_values:三维中z维度的值
            labels:每个参数更新轨迹的标签
            colors:每个轨迹的颜色
            interval:帧之间的延迟(以毫秒为单位)
            blit:是否优化绘图
        """
        self.fig = fig
        self.ax = ax
        self.xy_values = xy_values
        self.z_values = z_values

        frames = max(xy_value.shape[0] for xy_value in xy_values)

        self.lines = [ax.plot([], [], [], label=label, color=color, lw=2)[0]
                      for _, label, color in zip_longest(xy_values, labels, colors)]
        self.points = [ax.plot([], [], [], color=color, markeredgewidth=1, markeredgecolor='black', marker='o')[0]
                       for _, color in zip_longest(xy_values, colors)]
        # print(self.lines)
        super(Visualization3D, self).__init__(fig, self.animate, init_func=self.init_animation, frames=frames,
                                              interval=interval, blit=blit, **kwargs)

    def init_animation(self):
        # 数值初始化
        for line in self.lines:
            line.set_data_3d([], [], [])
        for point in self.points:
            point.set_data_3d([], [], [])
        return self.points + self.lines

    def animate(self, i):
        # 将x,y,z三个数据传入,绘制三维图像
        for line, xy_value, z_value in zip(self.lines, self.xy_values, self.z_values):
            line.set_data_3d(xy_value[:i, 0], xy_value[:i, 1], z_value[:i])
        for point, xy_value, z_value in zip(self.points, self.xy_values, self.z_values):
            point.set_data_3d([xy_value[i, 0]], [xy_value[i, 1]], [z_value[i]])

        return self.points + self.lines


def train_f(model, optimizer, x_init, epoch):
    x = x_init
    all_x = []
    losses = []
    for i in range(epoch):
        all_x.append(copy.deepcopy(x.numpy()))  # 浅拷贝 改为 深拷贝, 否则List的原值会被改变。 Edit by David 2022.12.4.
        loss = model(x)
        losses.append(loss)
        model.backward()
        optimizer.step()
        x = model.params['x']
    return torch.Tensor(np.array(all_x)), losses


# 构建5个模型,分别配备不同的优化器
model1 = OptimizedFunction3D()
opt_gd = SimpleBatchGD(init_lr=0.05, model=model1)

model2 = OptimizedFunction3D()
opt_adagrad = Adagrad(init_lr=0.05, model=model2, epsilon=1e-7)

model3 = OptimizedFunction3D()
opt_rmsprop = RMSprop(init_lr=0.05, model=model3, beta=0.9, epsilon=1e-7)

model4 = OptimizedFunction3D()
opt_momentum = Momentum(init_lr=0.05, model=model4, rho=0.9)

model5 = OptimizedFunction3D()
opt_adam = Adam(init_lr=0.05, model=model5, beta1=0.9, beta2=0.99, epsilon=1e-7)

models = [model5, model2, model3, model4, model1]
opts = [opt_adam, opt_adagrad, opt_rmsprop, opt_momentum, opt_gd]

x_all_opts = []
z_all_opts = []

# 使用不同优化器训练

for model, opt in zip(models, opts):
    x_init = torch.FloatTensor([0.00001, 0.5])
    x_one_opt, z_one_opt = train_f(model, opt, x_init, 100)  # epoch
    # 保存参数值
    x_all_opts.append(x_one_opt.numpy())
    z_all_opts.append(np.squeeze(z_one_opt))

# 使用numpy.meshgrid生成x1,x2矩阵,矩阵的每一行为[-3, 3],以0.1为间隔的数值
x1 = np.arange(-1, 2, 0.01)
x2 = np.arange(-1, 1, 0.05)
x1, x2 = np.meshgrid(x1, x2)
init_x = torch.Tensor(np.array([x1, x2]))

model = OptimizedFunction3D()

# 绘制 f_3d函数 的 三维图像
fig = plt.figure()
ax = plt.axes(projection='3d')
X = init_x[0].numpy()
Y = init_x[1].numpy()
Z = model(init_x).numpy()  # 改为 model(init_x).numpy() David 2022.12.4
surf = ax.plot_surface(X, Y, Z, edgecolor='grey', cmap=cm.coolwarm)
# fig.colorbar(surf, shrink=0.5, aspect=1)
ax.set_zlim(-3, 2)
ax.set_xlabel('x1')
ax.set_ylabel('x2')
ax.set_zlabel('f(x1,x2)')

labels = ['Adam', 'AdaGrad', 'RMSprop', 'Momentum', 'SGD']
colors = ['#8B0000', '#0000FF', '#000000', '#008B00', '#FF0000']

animator = Visualization3D(*x_all_opts, z_values=z_all_opts, labels=labels, colors=colors, fig=fig, ax=ax)
ax.legend(loc='upper right')

plt.show()
# animator.save('teaser' + '.gif', writer='imagemagick',fps=10) # 效果不好,估计被挡住了…… 有待进一步提高 Edit by David 2022.12.4
# save不好用,不费劲了,安装个软件做gif https://pc.qq.com/detail/13/detail_23913.html

【分析】:

       SGD陷入了鞍点,无法更新了。

        AdaGrad 方向很稳定,但是由于学习率衰减,更新速度越来越慢。

        RMSprop下降速度很快而且方向也很稳定。

        Momentum刚开始还不是特别快,更新方向也不是很准,但是往后在正确的方向上越来越快,跳出鞍点

        Adam下降速度比RMSprop稍微慢一点但速度也不慢,方向也很稳定 。

4 参考链接 

NNDL实验 优化算法3D轨迹 复现cs231经典动画_优化算法寻优过程3d-CSDN博客

【NNDL 作业】优化算法比较 增加 RMSprop、Nesterov_rmsprop+nesterov的python实现-CSDN博客

NNDL 作业11:优化算法比较_optimizers = ordereddict() optimizers["sgd"] = sgd-CSDN博客

深度学习优化算法比较_深度学习 常用优化算法比较-CSDN博客

Adam那么棒,为什么还对SGD念念不忘 (3)—— 优化算法的选择与使用策略 - 知乎

深度学习中的优化算法总结 - ZingpLiu - 博客园


网站公告

今日签到

点亮在社区的每一天
去签到